Research On Centralized Trust Management For Cognitive Radio Networks

Cognitive Radio Networks is composed of many cognitive users with sensing andlearning modules, which can access the spectrum dynamically. Some new attacks haveappeared because of the special features of the Cognitive Radio Networks, such ascollaboration between cognitive users, sensing ability, learning ability, self-adaptive andreconfigurable. Therefore, with more attention being paid to the study of CognitiveRadio Networks, security is becoming the key of whether it can be widely used. Inparticular, the reliability of the data is becoming the prerequisite of whether all thedecision-making processes are accurate and fair. As a complement of the cryptographicmeasures, trust and reputation management systems can counter the soft security threatseffectively and provide supervision of the behavior to improve the fairness androbustness of the network.Most of the current trust and reputation mechanisms in the Cognitive RadioNetworks are established to solve the problem of mistrust data in certain module of thecognition cycle, such as false data in the process of spectrum sensing and allocation ofchannels. However, the modules of the whole cognition cycle work together and interactwith each other. The security attribute of one module provides protection for themodules after. It is necessary to establish a trust management mechanism around theentire cognition cycle to provide reliable security assurance for dynamic spectrumaccess.This paper proposes a centralized trust management mechanism in CognitiveRadio Networks, which considers the characteristics of the network, the differentrequirements of security in each module in the cognition cycle and the overall securityrequirements of the network. Due to the authorized users and the cognitive users sharethe same spectrum resource, the trust management model is built with double trustrelationships, one is between the authorized user and the cognitive user, the other one isbetween the cognitive users. The trust management mechanism proposed collects thefresh trust evidence from the internal evidence and the external evidence, then updatesthe reputation of each user using the historical reputation and the fresh evidence, andthen finishes the trust assessment with the dynamic trust assessment mechanism.According to these results, reward mechanism will take related measure. The analysisshows that the trust management mechanism proposed in this paper can effectively detect the selfish and malicious behavior to make the network fair and reliable.